Design Patterns: An Essential Component of CS Curricula

Design Patterns:
An Essential Component of CS Curricula
Owen Astrachan, Geoffrey Berry
Landon Cox, Garrett Mitchener
Department of Computer Science
Duke University
[email protected]
NSF DUE-9554910, CCR-9702550
Duke University Computer Science : Patterns
NSF support for this project
DUE-9554910, Course, Curriculum, and Development
➤ Applied Apprenticeship Approach to CS2, interinstitutional collaboration: Duke, Appalachian State
University, North Carolina Central University
➤ Develop instructional frameworks for assignments, labs,
• apprentice learning: read, modify, write
• assignment re-use, develop frameworks that allow re-use, but
permit different assignments each semester
CCR-9702550, CAREER
➤ Using and Developing Design Patterns in Undergraudate
Computer Science Courses
➤ instructional material and curricular suggestions for
studying and using patterns (primarily first three courses)
Duke University Computer Science : Patterns
Explaining, Using, Developing Patterns
What is a pattern, why patterns are important
Why software infrastructure is essential in explaining,
understanding, and using patterns
Programming Patterns, Design Patterns, other kinds of
Incorporating patterns into courses, rethinking courses
Duke University Computer Science : Patterns
Overview of talk/discussion/interaction
Patterns, genesis, use in developing software
programming patterns, object-oriented design
before and after, class libraries, software cadavers,
pattern infrastructure
using patterns in courses, developing pattern-based
discussion, future work
Duke University Computer Science : Patterns
What is a pattern?
“… a three part rule, which expresses a relation between a
certain context, a problem, and a solution. The pattern is, in
short, at the same time a thing, … , and the rule which tells us
how to create that thing, and when we must create it.”
Christopher Alexander
factory, aka virtual constructor
➤ problem
delegate creation responsibility: expression tree nodes
➤ solution
createFoo() method returns aFoo, bFoo,...
➤ consequences potentially lots of subclassing, ...
more a recipe than a plan, micro-architecture, frameworks,
language idioms made abstract, less than a principle but more
than a heuristic
patterns capture important practice in a form that makes the
practice accessible
Duke University Computer Science : Patterns
Patterns are discovered, not invented
You encounter the same “pattern” in developing solutions to
programming or design problems
➤ develop the pattern into an appropriate form that makes it
accessible to others
➤ fit the pattern into a language of other, related patterns
Patterns transcend programming languages, but not (always)
programming paradigms
➤ OO folk started the patterns movement
➤ language idioms, programming templates, programming
patterns, case studies
Duke University Computer Science : Patterns
Pattern/Programming Interlude
3 3 5 5 7 8 8 8
Microsoft interview question (1998)
Dutch National Flag problem (1976)
Remove Zeros (AP 1987)
3 5 7 8
2 1 0 5 0 0 8 4
2 1 5 8 4
Quicksort partition (1961, 1986)
4 3 8 9 1 6 0 5
3 1 0 4 8 9 6 5
Run-length encoding (SIGCSE 1998)
11 3 5 3 2 6 2 6 5 3 5 3 5 3 10
Duke University Computer Science : Patterns
Solving (related) problems
Sometimes it is not clear that problems are related, or how
problems are related
Educators sometimes do not know what they know, so cannot
convey knowledge of how to solve problems to students
➤ often students don’t see programming problems as related,
or see them related by language features rather than by
higher-level features/dependencies
➤ it’s often difficult for students to appreciate why one
method of solving a problem is better without a context to
see the solution in force more than once
Using patterns can help make knowledge gleaned over many
years accessible to those new to the field
➤ patterns may be useful in connecting problems and
providing a framework for categorizing solutions
Duke University Computer Science : Patterns
One loop for linear structures
Algorithmically, a problem may seem to call for multiple
loops to match intuition on how control structures are used to
program a solution to the problem, but data is stored
sequentially, e.g., in an array or file. Programming based on
control leads to more problems than programming based on
Therefore, use the structure of the data to guide the
programmed solution: one loop for sequential data with
appropriately guarded conditionals to implement the control
Consequences: one loop really means loop according to
structure, do not add loops for control: what does the code
look like for run-length encoding example?
What about efficiency?
Duke University Computer Science : Patterns
One loop/linear: efficiency issues
Consider Quicksort partition example
➤ see code samples from different text books
➤ consider student developed Quicksort
• write from scratch? Get code from a book?
Quicksort in practice: Engineering a Sort Function, Bentley and
McIlroy, Software Practice and Experience, November 1993
• what about equal keys?
Consequence of pattern: possible that some efficiency may be
➤ when is efficiency important? Lessons for students?
➤ Make it run, make it right, make it fast, make it small
Duke University Computer Science : Patterns
Coding Pattern
➤ one loop for linear structures
➤ Sequential data, e.g., in an array or a file, must be
processed to perform some algorithmic task. At first it
may seem that multiple (nested) loops are needed, but
developing such loops correctly is often hard in practice.
➤ Let the structure of the data guide the coding solution. Use
one loop with guarded/if statements when processing onedimensional, linear/sequential data
➤ Code is simpler to reason about, facilitates develop of loop
invariants, possibly leads to (slightly?) less efficient code
Duke University Computer Science : Patterns
Simply Understood Code
People need to stare at code in order to understand it well enough to
feel secure making changes to it. Spending time switching from
window to window or scrolling up and down to see all the relevant
portions of a code fragment takes attention away from understanding
the code and gaining confidence to modify it.
People can more readily understand things that they can read in their
natural text reading order; for Western culture this is generally left to
right, top to bottom.
If code cannot be confidently understood, it will be accidentally
Therefore, Arrange the important parts of the code so it fits on one
page. Make that code understandable to a person reading it from top
to bottom. Do not require the code to be repeatedly scanned in order
to understand how data is used and how control moves about.
Duke University Computer Science : Patterns
OO Design Patterns
Assumption: the object-oriented paradigm is simply better
than the structured programming paradigm. OO is here to
stay, and will be used with increasing frequency in CS
courses, including (especially?) CS1 and CS2
➤ Consequence: it is hard, especially at first, to write and
develop code in an object-oriented way, especially for
those whose synapses have been seemingly hard-wired
from years of Fortran, Pascal, C, …
➤ Problem: most (all?) CS1 and CS2 texts do not show good
examples of OO programming and design
➤ Solution: design patterns can help with both problems, but
the classic software pattern literature (e.g., GOF) is not
easily digestible
Limit the number of patterns studied, use before and after
Duke University Computer Science : Patterns
Before and After
Problem: the “animal guessing game”
is it a mammal? Yes
does it have stripes? No
does it live in Africa? No
is it a moose? No
I give up, what animal were you thinking of? Platypus
do you want to play again? ...
Classic data structures/CS2 assignment
➤ uses binary trees, engaging (potentially) to students
➤ read/store from/to a file
➤ what lessons are learned from this assignment? What are
the goals in assigning it?
Duke University Computer Science : Patterns
Structured/Class-based approach
This solution (see code handout) is from Main and Savitch, Data
Structures and Other Objects using C++, A-W
➤ How are tree nodes created? What’s a more object-oriented
way of doing this?
➤ Where is the control for playing a game?
➤ Where would we add code for reading/saving from/to a file?
What programming lessons can we and our students extrapolate
from this exercise?
Each program we show to students, or ask students to write, is a
precious resource
Duke University Computer Science : Patterns
Using OO techniques/patterns
(my) Fundamental law of OO
programming and design
Ask not what you can do to an
object, ask what an object can
do to itself
Nodes ask questions, nodes
print themselves (to a file)
How to differentiate between
internal and external nodes?
➤ Inheritance hierarchy, use
abstract base class
Who creates nodes, the game?
➤ A node factory
➤ A game factory
Duke University Computer Science : Patterns
This diagram (and code) from Duke’s CS2
course, materials developed for AAA
project, see web for details
Before and After
(roughly) 300 lines of code for both,
OO version has 100 lines of .h/.cpp
for load and save
where is code added to
read a directory of data
files and give user a
choice of game?
Upside of OO
➤ transferable design
and practice,...
Downside of OO
➤ arguably more
Duke University Computer Science : Patterns
Software frameworks/cadavers
Experience with OO programming and design shows that
design patterns are useful
➤ where do we get the experience?
➤ how do we give experience to our students?
➤ what can we use to illustrate patterns in practice?
➤ what patterns should we emphasize, how, when?
Good design comes from experience, experience comes from bad design
Fred Brooks/Henry Petroski
Good design also comes from experiences with good design
➤ use several software cadavers to show patterns in practice
Duke University Computer Science : Patterns
Darwin, example of software cadaver
Originally developed at Stanford
for use in their courses
➤ Ported to Duke in Java and
C++ (qt widget toolkit)
➤ Used in COSEN, program for
women and minorities to
explain computer science
➤ beings inhabit a world, beings
can change species
➤ behavior is programmed using
DULL (Darwin Unstructured
Lattice Language)
➤ students implement parts of
Duke University Computer Science : Patterns
Factory pattern
Provide interfaces for different parts of the program, expect
students to implement the interface
➤ Students should not need to implement the entire program
at once; vary the assignment, get students to grow the
➤ Student implements class Creature, how is Creature used
in the rest of the program [that student may not write]?
• Students create a CreatureFactory that returns FooCreature,
which implements/subclasses Creature
Factory * fact = new CreatureFactory;
Creature * c = fact->makeCreature();
A factory is often implemented as a singleton, but avoid toomuch-too-soon syndrome
Duke University Computer Science : Patterns
When the creatures move, the world should show their
➤ when a program executes, the program view changes
➤ each observable (creature) notifies its observers (world
listener, program listener) when observable changes
➤ separate the model from the view, especially useful when
developing GUI programs, allows multiple views of the
same model
Use pattern in one course, show again in a subsequent
assignment/course, eventually students adopt/understand
➤ spiral approach to using patterns is essential for success
Duke University Computer Science : Patterns
Collections must support access to their elements, use a
separate class for access, e.g., supporting forward, backward,
sequential, random access
➤ iterators are essential in STL, Enumeration used in Java
1.1, Iterator used in 1.2
Use the Iterator pattern early in CS1 to hide platform-specific
code (e.g., in C++), to hide complicated code (e.g., I/O in Java),
and to introduce an important pattern
➤ WordStreamIterator in C++ and in Java
➤ Directory reading/processing classes in both languages
Internal iterators useful also, even in non OO languages
Duke University Computer Science : Patterns
Design patterns you shouldn’t miss
➤ useful in many contexts, see previous examples, integral to
both C++ and Java
➤ essential for developing OO programs/classes, e.g., create
iterator from a Java 1.2 List? list.iterator()
➤ essential in GUI/Widget programming, widgets contain
collections of other widgets
➤ replug-and-play, hide details
Observer/Observable, Publish/Subscribe, MVC
➤ separate the model from the view, smart updates
Duke University Computer Science : Patterns
Patterns discussed in other venues
Template pattern, function objects
➤ see Gelfand, Goodrich, Tamassia (SIGCSE 98)
Singleton, null-object, state pattern
➤ see Nguyen (SIGCSE 98)
➤ see Fell, Proulx, Rasala (SIGCSE 98)
Elementary programming patterns
➤ see Wallingford (SIGCSE 96)
➤ ChiliPloP (98)
Do we need to worry about what a pattern is?
➤ Growing literature, must contributors conform?
Duke University Computer Science : Patterns
Course and Curricular changes
Courses must become problem oriented rather than language
➤ texts should explain language and programming details in
a context, one that is rich in computer science, software
design (or engineering), and intrigue
Patterns may be part of a new way to organize courses and
➤ one course is not enough, two courses are not enough,
think long-term, develop material over a three-course
Provide pattern-languages/catalogues, facilitate use and
application of patterns
➤ cannot throw GOF (or equivalent) at beginners
Duke University Computer Science : Patterns